Location: south-facing roof in Columbia River Gorge, 65 kilometers east of Portland, OR.  

The inverter can make switching decisions such as: 1) if no exterior power (e.g., downed transmission lines), route from batteries to house demand; 2) if house demand is less than solar-based input, charge batteries; 3) if 2) and batteries are charged, send to the exterior power grid (turn meter backwards).  OK - this is conventional stuff nowadays.

Also, water-to-air heat pumps are not only well-known, but the market is growing at an encouraging rate - encouraging because this is demonstrated to be the most efficient conventional approach to space-heating/cooling.  Basically, it uses the well-known refrigeration cycle of expansion/compression to concentrate heat in one region of the machine and to remove heat from another region.

For those who don't know about the so-called geothermal heat pump system, it is typically based on pipes set about 1.7 meters deep in the ground, where soil temperature stays fairly stable at close to 10 degrees C in the temperate zones of the world.  In Winter the refrigeration cycle is designed such that the heat pump pulls out some of the heat inherent in 10 degree water, sending, say, 5 degree water back into the pipes in the ground.  The length of the piping system is calculated to permit the water to equilibrate at the ground temperature before returning to the heat pump.  In Summer the system is valved such that the system reverses direction in terms of heat flow - the heated water goes out to the pipes in the ground.  The piping systems are typically quite long, but the extent of the trenching can be reduced by digging wider trenches and looping the pipe as it is laid.

Another less-used system (that is also becoming more common) is to use black rubber pads with small channels fabricated into the length of the pads, manifolded into pipes running width-wise at either end of the pads, to capture solar-based heat in water flowing through these channels.  In the U.S. swimming pools are sometimes warmed in the Spring and Fall by this method.  Occasionally, these pads are used in conjunction with storage tanks to provide warm/hot water for 'hydronic' heating of floors - water-carrying tubes laid in thick mortar beds under tiles, for instance.

The idea/experiment here is to combine the heating via the black pads with a water-to-air heat pump.  One ½ hp pump will drive the water from the storage tanks through the pads on the roof and back into the tanks.  A second pump will take water from the tanks to the heat pump, when a house-interior thermostat demands hot (or cold) air.

My son helped me to install the water-heating pads on my roof two weekends ago.  The last pieces were the water storage tanks, which are sitting in my driveway now.  I'm getting ready to install my tanks, pumps, and heat pump in the next few weeks.  (I want to get them up soon, so that I can start collecting temperature data vs. ambient conditions in Winter.)  After that I'll install my solar modules, grid-tie inverter, and batteries.  (Plus I will buy another 2 kw-capability of photovoltaic modules by the end of the year.)

Conditions in my neighborhood are anything but ideal. Insolation runs at about 60% of the high-prairie region just east of my county. Insolation data for this area says that I should just about cover the southern half of my roof with panels to supply about the same kwh-equivalent that I currently consume in electric resistance heating - if capture is successful.

Days here are frequently windy, and a friend - Ormond O. - predicts that the wind will actually work to pull heat out of the system - or at least to reduce the capture of the potential solar-based heat. We'll see.  If glazed systems are needed to overcome this effect, at least I won't have much invested in the rubber ones (they're quite cheap).  In addition my roof is only sloped at about 10 degrees from the horizontal; and, since we are above the 45th parallel, best angle would probably be something like 60 degrees in Winter.

Couple of interesting wrinkles to consider:
1)The tanks can be charged from rainwater on the roof;
2)Since my garden is one "floor" beneath my garage, I can water the garden with rainwater via the storage tanks;
3)In Fall, Winter, or Spring - when the rainwater is warmer than the water in the tanks, the rainwater can be used to supplant the tank water, raising the temperature and, thus, providing heat;
4)In Winter the water temperature from our city system comes in at about 10 C.  The city system is gravity-pressurized.  In the case of long-term electrical failure, the heat pump water system can be recharged from city water, and the heat pump can be run for more than one month on the batteries, assuming high-charge state initially;
5)In Summer the pump that lifts the water to the pads would be turned on at, say, midnight and off at, say, 5:00 AM.  Idea would be to radiate heat away from the pads during the coolest part of the night.  This would be for use in supporting the 'air conditioner' cycle.  (We see 35 C or above about one to two weeks per year, so air conditioning would be nice during that period.)

So - as I say - this is an experiment.  Comparing temperature changes to the various ambient conditions should provide some ability to predict viability - although viability may entail moving 80 km to the East of here.